Aurora for PostgreSQL 到 SelectDB

选择对端数据库：

SelectDB

数据链路

基本功能

功能	说明
Schema Migration	If the target schema does not exist, BladePipe will automatically generate and execute CREATE statements based on the source metadata and the mapping rule. This feature is not applicable to the tables without primary keys.
Full Data Migration	Migrate data by sequentially scanning data in tables and writing it in batches to the target database. This feature is not applicable to the tables without primary keys.
Incremental Data Sync	Sync of common DML like INSERT, UPDATE, DELETE is supported. This feature is not applicable to the tables without primary keys.
Data Verification and Correction	Verify all existing data. Optionally, you can correct the inconsistent data based on verification results. Scheduled DataTasks are supported. For more information, see Create Verification and Correction DataJob.
Subscription Modification	Add, delete, or modify the subscribed tables with support for historical data migration. For more information, see Modify Subscription.
Position Resetting	Reset positions by file position or timestamp. Allow re-consumption of incremental data logs in a past period or since a specific Binlog file and position.
Table Name Mapping	Support the mapping rules, namely, keeping the name the same as that in Source, converting the text to lowercase, converting the text to uppercase, truncating the name by "_digit" suffix.
DDL Sync	PostgreSQL DDL sync is realized by triggers. The user should have the permissions on triggers and tables. For more information, see Permissions Required for PostgreSQL ALTER TABLE ADD COLUMN, MODIFY COLUMN, RENAME COLUMN, DROP COLUMN RENAME TABLE CREATE TABLE (whole database sync)
Metadata Retrieval	Retrieve the target metadata with filtering conditions or target primary keys set from the source table.

高级功能

功能	说明
全量前清空目标数据	运行全量任务前清除老数据，包括重跑任务、定时全量迁移都会触发此能力
重建目标表	运行全量任务前重建目标表，包括重跑任务、定时全量迁移都会触发此能力
Stream Load 数据写入	采用 Stream Load 到 Doris / SelectDB be 写入数据, 默认攒批写入，可动态调节刷出数据节奏和批次大小
0 值时间处理	支持将 0 值时间设置成不同类型的值，防止写入对端报错
Online DDL Compatibility	Supports synchronization for GH-OST, PT-OSC, Aliyun DMS Online DDL, and PT_GHOST.
Setting Data Partitions	When creating a DataJob, specify partition definitions at the table level (static or dynamic). Automatically add these partition definitions during schema migration.
Scheduled Full Data Migration	For more information, see Create Scheduled Full Data DataJob.
Custom Code	For more information, see Custom Code Processing, Debug Custom Code and Logging in Custom Code.
Adding Virtual Columns	Support adding custom virtual columns with fixed values, such as region, ID, etc.
Setting Target Primary Key	Change the primary key to another field to facilitate data aggregation and other operations.
Data Filtering Conditions	Support data filtering using WHERE conditions, with SQL-92 as the SQL language. For more information, see Data Filtering.

限制和注意点

限制项	说明
对端表类型	仅支持 唯一键模型(Unique)
源端表类型	不支持 无主键表 迁移同步
数据类型	不支持 BINARY, BLOB 等二进制类型
增量写入冲突策略	Stream Load 写入以主键进行整行替换

链路FAQ

• Doris / SelectDB 目标端写入报错

---

源端数据源

前置条件

条件	说明
账号权限	需要权限如下(以自建数据库为例): GRANT ALL PRIVILEGES ON DATABASE 同步库 TO 同步账号（或同步库 information_schema 中所有视图的 SELECT 权限和需要同步表、索引、约束的 SELECT 权限) ALTER USER 同步账号 REPLICATION
增量同步准备	准备动作按如下步骤进行: 修改 postgresql.conf, 设置 wal_level=logical 和 wal_log_hints = on 修改 pg_hba.conf, 设置 host replication 同步账号 CIDR网段 md5 , host 同步库 同步账号 CIDR网段 md5, host postgres 同步账号 CIDR网段 md5 重启 PostgreSQL
网络准备	迁移同步节点（sidecar）可连接 PostgreSQL 标准交互接口（如 5432）

任务参数

参数名称	说明
fullFetchSize	全量扫描数据设置的 fetch size
eventStoreSize	缓存解析完毕的增量事件缓存大小
ignoreGisSRID	解析 GIS 数据类型时是否忽略 SRID
defaultGisSRID	设置 GIS 数据类型的 SRID

Tips: 通用参数配置请参考 通用参数及功能

---

目标端数据源

前置条件

条件	说明
账号权限	具备 SELECT, DDL 权限(可选)
网络准备	迁移同步节点（sidecar）可连接 Doris / SelectDB FE QueryPort 和 FE/BE HttpPort

任务参数

参数名称	说明
host	MySQL 协议交互链接，对应 Doris / SelectDB FE QueryPort
httpHost	Doris stream load 链接，对应 Doris / SelectDB FE/BE HttpPort
totalDataInMemMb	攒批写入，内存中最大数据容量，超过此容量或超过 asyncFlushIntervalSec 则刷出数据到写入队列
asyncFlushIntervalSec	攒批写入，等待刷出的间隔时间，超过此时间或超过 totalDataInMemMb 则刷出数据到写入队列
flushBatchMb	单表最大攒批容量，超过此容量则刷出数据到写入队列
realFlushPauseSec	使用 stream load 刷出数据到 Doris / SelectDB 的等待时间，0 则不等待
soTimeoutSec	在 QueryPort 或 HttpPort 执行操作时 tcp 超时链接(so_timeout)
enableTimeZoneProcess	是否对时间字段进行时区转换
timezone	目标端 Doris/SelectDB 时区，例如 +08:00 Asia/Shanghai America/New_York
maxInSizePerQuery	校验任务中，对端单次查询的最大 IN 条件值数量，大于该值会自动拆分多次查询

Tips: 通用参数配置请参考 通用参数及功能